Ali Hortacsu - US grants

This research explores a number of facets of the competitive market structure in the U.S. mutual fund industry, specifically with regard to the nature of mutual fund firms' strategic decisions. While a considerable amount of research in financial economics has been devoted to mutual funds, its primary focus has been on aspects of mutual fund operation in isolation from other industry firms: portfolio construction, evaluation of fund performance vis-a-vis benchmark portfolios, and the agency relationship between shareholders and the fund manager. Previous research has not delved greatly into the strategic decisions of industry firms. The project fills this gap and builds on the recent increase of interest among financial economists in the mutual fund industry's competitive structure. It explicitly models several aspects of competitive behavior in the industry, offering testable and excludable hypotheses. It also investigates elements of industry competition that have not been studied in depth.

Specific questions that this project addresses include the following: (1) Do investors face
significant search costs/informational frictions when choosing funds? (2) What non-risk or return attributes characterize investors' demand for mutual funds? (3) How do fund companies set
management fees? (4) What non-price aspects of competition exist in the mutual fund industry?
(5) Is the existing competitive structure of the industry optimal from a social welfare standpoint?
The project takes several approaches to this aim. Specific topics are investigated in sector case studies that control for confounding influences on outcomes and highlight the competitive
decision(s) of interest. Further, the research involves modeling market interactions and the
estimation of the models' structural parameters using market data. This allows the evaluation of actual, anticipated, or proposed market changes on industry structure and consumer and social welfare, facilitating policy counterfactuals and hypothesis testing.

This research offers potentially substantial broader impact. The mutual fund industry is an important part of the domestic financial system, and as such has been of great interest to popular media, regulators, and policymakers. In 2001, for example, over half of U.S. households had mutual fund holdings, and various institutional fiduciaries held over $3 trillion in assets. The industry also plays a major role in households' retirement plans: roughly one-third of mutual fund assets are held in retirement savings vehicles. Given its size and ubiquity, the industry's competitive market structure and the strategic decisions shaping it directly affect not only the fortunes of individual funds and fund families, but consumer welfare as well. Improved
understanding of these decisions and how they respond to real and prospective changes in the
competitive environment will undoubtedly yield useful insights to researchers and policymakers.
Furthermore, the project's research synthesizes elements of the relevant finance literature with
industrial organization research on consumer search behavior, product differentiation, producer entry, and oligopolistic price competition. As such, it holds considerable potential to spur additional interchange of theoretical and empirical ideas and methodologies between the industrial organization and finance literatures. The research program will emphasize the development of novel empirical methods and probe the applicability of current analytical tools of the industrial organization literature to the financial economics arena.

Abstract:
This CAREER project consists of research and educational projects centered around the empirical analysis of auction and matching markets. The development of the Internet and surrounding computing and communication technologies have led to an explosion in the use of auctions or other mechanisms to set prices and match buyers and sellers in markets for many diverse goods and services. Although there have been considerable advances in economic theories of how such market mechanisms work, empirical work so far has been limited by the availability of data and the realism of economic models utilized to depict real-world environments. The research component of the project will conduct in-depth empirical studies of three distinct but important market settings where data is available and the gap between theory and empirics is surmountable. The first research project in the agenda will focus on the ERCOT (Electricity Reliability Council of Texas) electricity balancing market. This is a market where electricity generators submit hourly supply bids to increase or decrease production to meet short-term fluctuations in electricity demand. Utilizing very detailed data on the bidding behavior of generation firms, along with engineering estimates of their marginal costs of producing electricity, the study will test economic theories of strategic behavior by the firms, with a focus on assessing the extent to which firms utilize market power, and how this impacts the productive efficiency achieved by a market that serves 20 million electricity users in Texas. The second research project studies auctions conducted by the Bank of Canada to sell Government of Canada bills and bonds. Again utilizing a very detailed bidder-level data set, the study will focus on analyzing the impact of restricting participation to these auctions to a set of certified government securities dealers. In particular, the study will investigate the complex interactions between securities dealers who are allowed to bid in the auction directly, and non-dealer investors who can only place bids through securities dealers. The study also aims to quantify the impact of proposed mergers between securities dealers on the public debt financing costs of the Government of Canada. The third research project shifts attention to a somewhat more controversial, but equally interesting market setting, that of Internet matchmaking. The aim of the project is to develop empirical methods to infer a diverse group of individuals' preferences over their potential match partners, based on data on the search and matching behavior of users on an online dating site. Based on our estimates of these preference patterns, we will then compare the efficiency of the particular matching protocol utilized on this site to the efficiency of matches obtained using theoretically motivated matching protocols, quantifying potential gains to improved market design. The educational component of the CAREER project aims to teach future generations of "market designers" how to utilize the rapidly developing toolbox of theoretical and empirical tools to analyze and formulate policy in many other important markets. Towards this aim, the PI will develop and teach a new graduate course on applied market design. To further undergraduate education in this rapidly advancing field, the PI will undertake an interdisciplinary collaboration with the University of Chicago computer science department to develop an online auction environment. Classroom experiments and demonstrations on this online environment will be integrated into the curriculum of undergraduate microeconomics and industrial organization courses taught by the PI to allow students to gain a working knowledge of real-world auction and pricing mechanisms

Neoclassical economic theory underlies almost all modern economic policy analyses. Nevertheless, several important deviations from standard assumptions of the theory have been robustly documented in laboratory experiments on economic decision-making. Two such important behavioral deviations are the "endowment effect" and the prevalent display of altruism/social preferences. A multidisciplinary team of researchers from economics, psychology, and neurology will examine the sources of these experimental findings. In this research, the economic decision-making experiments - spanning both laboratory and field treatments - used to detect these deviations will be complemented by neuroimaging techniques to provide insights into the neural mechanisms underlying the behavioral responses. An important hypothesis that will be tested in the context of the endowment effect is whether the effect survives in competitive trading settings. To test this hypothesis, a subject pool with varying degrees of experience as traders in real-world markets will be utilized. Neuroimaging techniques will be used to examine the neural responses of experienced versus inexperienced subjects. The experimental subjects will also be tracked over the course of several years to examine the persistence of behavioral patterns, and whether market experience attained during the survey period is reflected in neural activation patterns. In order to isolate the sources of social preferences/altruism displayed in laboratory games, an experimental design with variations in giving contexts will be used to test among competing explanations for the observed behavioral responses. Neuroimaging will play a crucial role in discriminating between alternative hypotheses and predicting behavior in novel situations. To collect neural measurements with high spatial and temporal accuracy, both fMRI and event-related electroencephalography (ER-EEG or ERP) techniques will be utilized, allowing a more complete understanding of the gains achieved from combining these two measurement modalities. The research is expected to have broader impact through several channels. The economic decision-making contexts studied in the research are ubiquitous in daily life. Assessing the effects of market experience and context variations on decision-making are necessary steps to extend the predictive domain of behavioral theories beyond the laboratory. The multidisciplinary nature of the project is expected to promote the flow of knowledge across the economic, psychology, and neurology/neuroscience disciplines, and will produce trainees (at the post-doc, graduate and undergraduate levels) who will further the communication between these fields.

This award funds research on how economic shocks move through an entire economy via network links between individual firms.

A modern economy is comprised of intricately linked production units. The structure of this production network is one of the key determinants of how shocks are transmitted throughout the economy. Understanding the structure of the production network and how it affects the propagation of economic shocks can inform government policy regarding how to prepare for and recover from adverse shocks, such as natural disasters, energy price spikes, and financial shocks such as the bankruptcy of key suppliers and customers.

The network structure of production and its effect on shock propagation has been studied in previous research, but only at the sectoral level using input-output tables. The first project will empirically characterize the inter-firm production network in the U.S., mapping the individual supplier-customer links in the production economy. The investigator develops a model of network formation to fit key facts about the network. That model is then used to simulate how shocks to individual firms propagate through the supply chain and, potentially, to the whole economy.

The second project will study intra-firm linkages between production plants owned by multi-plant firms. Understanding why firms expand into vertically related businesses can provide us with insights regarding how managerial/entrepreneurial talent diffuses through the economy, and also potentially inform policy regarding vertical acquisitions/mergers. The project uncovers and aims to explain an empirical puzzle: there are very few internal shipments between seemingly vertically related plants of multi-plant firms. This suggests vertical integration is not necessarily about facilitating the flow of physical inputs between productive units, but rather intangible inputs, such as management practices and R&D output/manufacturing know-how.

The third project in the proposal was motivated by U.S. Treasury's Auto Warranty Commitment program in 2009, which provided a government back-stop against warranties issued by GM and Chrysler. The investigator studies the linkage between the financial health of warranty-issuing durable goods manufacturers (such as auto manufacturers) and consumer expectations regarding the firm's viability. After establishing the existence and magnitude of this link in the U.S. auto industry using very detailed data on used car auctions, the potential for fundamental instabilities in the U.S. auto manufacturing sector due to multiple equilibria/bank-run type dynamics is studied through an empirically calibrated model. The model is then used to assess the effectiveness of government policies (such as warranty back-stop programs) to thwart such dynamics.

Market design--economic engineering--has been successful in the redesign of real-world "two-sided matching markets," including the matching of doctors to residencies and the assignment of students to schools. A recent and growing literature has shown that the methods and results of two-sided matching theory extend to substantially more general settings, including "trading networks" in which agents negotiate over contracts.

Unlike two-sided matching markets, trading networks feature intermediaries, agents who both buy and sell in the market. These intermediaries may be intermediate producers (as in most economic models of supply chains) or may simply facilitate trade (e.g., through provision of information). Although the presence of intermediaries is the defining distinction between the matching-in-networks model and the previous two-sided frameworks, the analysis of intermediaries in these matching models has been limited.

This research will further our understanding of the behavior and impact of market intermediaries. Theoretical work using the matching-in-networks framework will be supplemented with detailed empirical analyses using plant-level microdata on production networks, in order to gain insight into the magnitudes of intermediaries' effects on welfare. Simulated counterfactual models will be developed when theory does not provide sharp predictions.

This work will inform and support new applications of market design approaches to real-world intermediated matching markets. One focus will be understanding how "local" perturbations (e.g., intermediary entrance and exit) affect the "global" structure of networks; another will be the identification and reduction of rent seeking by intermediaries.

In addition to enabling novel work in applied market design, this research will promote joint work between computer scientists and economists working in the theory and practice of market design.

This award funds research on the role of information in driving consumer choices, with special attention to retail banking and electricity industries.

Traditional economic models of consumer behavior typically assume that consumers are fully informed about their alternatives. However, the observation of persistent price dispersion in many product markets suggests that significant search and switching frictions are present in many markets. Such frictions may impede forces of competition from achieving efficient outcomes, and understanding the relevance and magnitude of these frictions is key to assessing the necessity and impact of regulatory interventions.

To fully understand and quantify the role played by these frictions, we need detailed information on consumer choices, and especially the choice-making process. Therefore, the research project utilizes detailed data on individual consumers' choices, incorporates explicit models of search/switching costs and allows for limited information about alternatives.

With its 14 trillion dollars of assets, about 7,000 banks and more than 80,000 bank branches, the U.S. retail banking sector is an extremely important portion of the retail economy as well as the financial system. Understanding how consumers decide on their retail banking provider and whether they face search and switching frictions is a crucial input into unpacking the drivers of competition and innovation in this important sector of the economy. The project on retail banking utilizes a unique data set with information on the entire purchase funnel of consumers' shopping process: (1) awareness, (2) consideration, and, finally, (3) purchase. The data set is among the first to provide detailed data on all stages of the purchase funnel, as well as being the first of its kind to study retail banking, a very important sector of the economy that has not been frequently scrutinized through the lens of modern empirical industrial organization. The data set is also complemented with information on advertising and product characteristics to estimate a very detailed model of shopping and decision-making at the individual consumer level. The data and the model allows the researchers to assess (1) what the magnitude of search and switching costs are that rationalize consumer behavior, and (2) whether advertising is primarily effective in increasing awareness or driving choice conditional on awareness.

The PI will also develop similar empirical methods to study consumer choices in deregulated retail electricity markets, which are in operation in 24 states. Insights from this study as well as the utilization of the empirical methods developed here in other electricity market contexts can aid market monitors and regulators to devise policies to further the competitiveness and efficiency of these markets. The project will utilize very detailed subscriber level data from Texas' residential electricity choice program. The research team will develop an econometric model that allows for search and switching based choice frictions. The model will be used to assess whether such choice frictions can explain the evolution of market shares and prices of the incumbent providers vs. entrants, and how interventions such as informational campaigns to reduce search/switching costs may affect market structure and consumer welfare.

Communication in a foreign language has been shown to strongly affect decision making. As communication and transportation technologies have increasingly "globalized" the world, more people communicate with one another using languages other than their native language, whether in business, politics, or casually. This interdisciplinary research project will use a variety of methods to evaluate the consequences of using a foreign language on individuals and society in a range of different domains. The investigators will assess the effects of foreign languages on strategic risk taking, auction market efficiency, public opinion surveys, creative problem solving, scientific discovery, invention, and entrepreneurship. The project will enhance basic understanding of the ways that language changes thought and behavior by providing new insights into the pervasive impact of language on intellectual activity, including mechanisms of mind, brain, and social institutions. As such, the project will make intellectual contributions across a range of the disciplines, including psychology, economics, political science, and sociology. The project also will have far-reaching social implications. For example, given that about one-half of the internal medicine residents in U.S. hospitals are foreign-born, health-care organizations can more precisely anticipate risk-taking trends among physicians depending on whether or not they use a foreign language professionally. Intelligence communities, diplomats, and businesses will be able to apply project findings strategically in negotiation and problem solving situations with foreign entities. The project also will help address central societal issues by providing new perspectives with which to assess the value of immigration and the innovation that accompanies it. With respect to education, the project will provide tools to increase the effectiveness of education and training for bilingual immigrant students and to better anticipate the benefits of bilingual education on performance in science, technology, engineering, and mathematics.

This project will focus on the ways through which language changes thought and behavior. Theories of individual and institutional behavior tend to be based strictly within disciplines and typically ignore the role of language. This project will transform these approaches by broaching the divide between individual decision makers and institutional behavior. The investigators will employ a novel interdisciplinary approach that involves collaboration among researchers from psychology, economics, sociology, and political science. They will use experimental methods and surveys to collect behavioral, physiological, and brain activity measures, comparing tasks performed in native and foreign languages. They also will engage in a large-scale, computational investigation of millions of publications and patents to evaluate the consequences of using a foreign language on innovations in science and business. This project is supported through the NSF Interdisciplinary Behavioral and Social Sciences Research (IBSS) competition.

The economies of many developing countries have experienced huge transformations over the past 20 years, but a major cost of development has been air and water pollution. The fact that pollution remains a central problem suggests that it may be very costly to mitigate pollution, in terms of lost productivity and revenues. An opposing view is the "Porter hypothesis" that postulates that environmental policies lead to greater productivity. The current empirical evidence on these tradeoffs is both limited and ambiguous, with some studies finding a substantial productivity cost of pollution mitigation and others finding evidence in favor of the Porter hypothesis. This research conducts a detailed empirical evaluation of whether policies in developing countries to lower pollution have been successful. The investigators also quantify the productivity and distributional costs of these policies. This research investigates these questions for both the manufacturing and power generation sectors.

This research makes a significant contribution to intellectual merit, through its use of novel data, its innovative identification, and its use of frontier methods. This research is among the first to combine firm-level panel datasets on industrial production, power plant investment in environmental mitigation technologies, and air and water pollution discharges, in developing or developed countries. The investigators develop new sources of identification to understand pollution in developing countries by focusing on panel variation in environmental discharge fees and pollution reduction mandates. Using the detailed data, the investigators also evaluate whether pollution mitigation policies cause firms to invest in new mitigation technologies or simply reduce their output.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.